better IsAlmostSimple support in AllPrimitiveGroups

[ThomasBreuer]

Up to now, the groups had to be constructed and were checked. Now we use that IsAlmostSimple is determined by ONanScottType.

The IsAlmostSimpleGroup flag is now set in the groups returned by PrimitiveGroup.

(The primgrp package does not say which functions have special support in AllPrimitiveGroups, but is_almost_simple is listed in the Oscar documentation.)

[codecov]

Codecov Report

❌ Patch coverage is 40.00000% with 3 lines in your changes missing coverage. Please review.
✅ Project coverage is 99.13%. Comparing base (dbae2c0) to head (ee7d23d).
⚠️ Report is 2 commits behind head on master.

Files with missing lines	Patch %	Lines
lib/primitiv.gi	40.00%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master      #79      +/-   ##
==========================================
+ Coverage   99.11%   99.13%   +0.02%     
==========================================
  Files          46       46              
  Lines       38324    38327       +3     
==========================================
+ Hits        37984    37995      +11     
+ Misses        340      332       -8     

Files with missing lines	Coverage Δ
lib/primitiv.gd	100.00% <ø> (ø)
lib/primitiv.gi	59.75% <40.00%> (+1.61%)	⬆️

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[fingolfin]

Thanks @ThomasBreuer! I let Claude have a closer look, and it pointed out three issues:

1. Documentation: lib/primitiv.gd:166-172 does list which properties come from the database. IsAlmostSimpleGroup should be added there.
2. CHANGES.md should be updated with this
3. "values" can also be lists, which this does not support

AllPrimitiveGroups(NrMovedPoints,12,IsAlmostSimple,[true])       ->  0  (want 6)
AllPrimitiveGroups(NrMovedPoints,12,IsAlmostSimple,[true,false]) ->  0  (want 6)
AllPrimitiveGroups(NrMovedPoints,12,IsSimple,[true,false])       ->  6  (correct)

Wrong answer, no error. STGSelFunc already dispatches on true/false/
function/list, so delegating to it fixes all cases and removes both the
four-way branch and the hand-maintained type blacklist:

elif a=IsAlmostSimpleGroup or a=IsAlmostSimple then
  # for primitive groups, almost simple means O'Nan-Scott type 2
  bfunc:= b;
  b:= val -> STGSelFunc(val = "2", bfunc);
  gut[i]:=Filtered(gut[i],j->STGSelFunc(PRIMGrp(i,j)[4],b));

[ThomasBreuer]

Just for the record:
I do not agree with Claude's suggestion for 3.

[fingolfin]

You wrote "for the record" that you don't agree with Claude's suggestion for 3. It was indeed needlessly complicated, this version already does it, and is very similar to the code right above and below this case:

Suggested change

	elif a=IsAlmostSimpleGroup or a=IsAlmostSimple then
	# for primitive groups, almost simple means O'Nan-Scott type 2
	if IsFunction(b) then
	# does not really make sense but is allowed
	gut[i]:=Filtered(gut[i],j->b(PRIMGrp(i,j)[4]="2"));
	else
	if IsBool(b) then
	b:= [b];
	fi;
	if IsList(b) then
	if true in b and false in b then
	# nothing is to do
	elif true in b then
	gut[i]:=Filtered(gut[i],j->PRIMGrp(i,j)[4]="2");
	elif false in b then
	gut[i]:=Filtered(gut[i],j->PRIMGrp(i,j)[4]<>"2");
	else
	gut[i]:=[];
	fi;
	else
	gut[i]:=[];
	fi;
	fi;
	elif a=IsAlmostSimpleGroup or a=IsAlmostSimple then
	# for primitive groups, almost simple means O'Nan-Scott type 2
	gut[i]:=Filtered(gut[i],j->STGSelFunc(PRIMGrp(i,j)[4]="2",b));

So, also for the record: what are your concerns with that, if there are any?

My main reason for preferring the version with STGSelFunc is that by having these checks similar, it is much easier for me to trust that they behave consistently. So the code becomes easier to maintain.

But I am happy to be convinced otherwise, I just haven't heard the arguments yet :-)

[ThomasBreuer]

O.k., apologies to Claude.